7/28/2019 Wp 12190

1/61

WP/12/190

Successful Austerity in the United States,Europe and Japan

Nicoletta Batini, Giovanni Callegari and Giovanni Melina

7/28/2019 Wp 12190

2/61

2012 International Monetary Fund WP/12/190

IMF Working Paper

EUR

Successful Austerity in the United States, Europe and Japan1

Prepared by Nicoletta Batini*, Giovanni Callegari** and Giovanni Melina***

Authorized for distribution by Peter Doyle

July 2012

Abstract

The output effects of 2009 fiscal expansions have been hotly debated. But the discussion of

fiscal multipliers is even more relevant now that several European countries have had to quickly

retract their stimulus measures in an effort to regain market confidence. Using regime-switching

VARs we estimate the impact of fiscal adjustment on the United States, Europe and Japan allowing

for fiscal multipliers to vary across recessions and booms. We also estimate ex ante probabilities of

recessions derived in association with different-sized and different types of consolidation shocks

(expenditure- versus tax-based). We use these estimates to understand how consolidations should be

designed to be most effective in terms of permanently and rapidly reducing a countrys debt-to-GDPratio. The main finding is that smooth and gradual consolidations are to be preferred to frontloaded or

aggressive consolidations, especially for economies in recession facing high risk premia on public

debt, because sheltering growth is key to the success of fiscal consolidation in these cases.

JEL Classification Numbers: E62; F41; H30; H63

Keywords: Fiscal consolidation, fiscal multipliers, growth-friendly fiscal policy

(*) Corresponding Authors E-Mail Address:NBatini@imf.org;giovanni.callegari@gmail.com;

giovanni.melina@gmail.com.

1 We are grateful to Anja Baum, Fabio Canova, Peter Doyle, Paul Levine, Anke Weber and seminar participants

at the International Monetay Fund and University of Surrey for useful comments. We thank Nathan Balke for

sharing his RATS codes and Thomas Warmedinger for sharing his debt simulation tool, which forms the basis

of our debt simulations. Brian Cho provided very valuable research assistance. We are also indebted to Bruno

Chiarini for sharing Italian accrual-based fiscal data and to Anke Weber for sharing Japanese revenue data.

Giovanni Melina acknowledges financial support from ESRC project RES-062-23-2451. Work on this paper

was done while Callegari was an Economist at the International Monetary Fund.

This Working Paper should not be reported as representing the views of the IMF.

The views expressed in this Working Paper are those of the author(s) and do not necessarily representthose of the IMF or IMF policy. Working Papers describe research in progress by the author(s) and are

published to elicit comments and to further debate.

7/28/2019 Wp 12190

3/61

Contents Page

Abstract ......................................................................................................................................1I. Introduction ............................................................................................................................4II. Literature Review ..................................................................................................................9

A. Empirical Literature on The Effects of Fiscal Shocks ..............................................9B. The Effects of Fiscal Policy: What do Models Say? ...............................................12C. Fiscal Policy Under Special Conditions ..................................................................15

III. Econometric Methodology.................................................................................................16 A. The Model ...............................................................................................................16B. Testing for the TVAR Structure ..............................................................................18C. Non-Linear Impulse Response Functions ...............................................................19D. Regime-Dependent Fiscal Multipliers ....................................................................19E. Conditional Probabilities of a Recession .................................................................20

IV. Results................................................................................................................................20A. Non-Linear IRFS.....................................................................................................21B. Fiscal Multipliers.....................................................................................................23C. Conditional Probabilities of a Recession Following a Fiscal Shock .......................25

V. What Does This Mean in Practice? .....................................................................................26A. Model Used for The Debt Simulations ...................................................................27 B. Debt simulations ......................................................................................................30

VI. Concluding Remarks .........................................................................................................32VII. References ........................................................................................................................33VIII. Data Appendix ................................................................................................................58

A. Euro Area (1985 Q1 2009 Q4):.............................................................................58B. France (1970 Q1 2010 Q4): .................................................................................58C. Italy (1981 Q1 2007 Q4): ......................................................................................59D. Japan (1981Q1 2009 Q4): .....................................................................................59E. United States (1975 Q1 2010 Q2): ........................................................................60

7/28/2019 Wp 12190

4/61

3

Tables

1. Tests for threshold VAR ..........................................................................................542. Cumulative Fiscal Multipliers ..................................................................................553. Conditional Probabilities of a Recession (Expenditure Cut) ...................................56

4. Conditional Probabilities of a Recession (Tax Hike) ..............................................57Figures

1. Average Responses to Shocks to Government SpendingEuro Area ..................................382. Average Responses to Shocks to Government SpendingFrance .......................................393. Average Responses to Shocks to Government SpendingItaly ...........................................404. Average Responses to Shocks to Government SpendingJapan .........................................41 5. Average Responses to Shocks to Government SpendingUnited States .............................426. Average Responses to Shocks to Net TaxesEuro Area .....................................................437. Average Responses to Shocks to Net TaxesFrance ...........................................................44 8. Average Responses to Shocks to Net TaxesItaly ...............................................................459. Average Responses to Shocks to Net TaxesJapan .............................................................4610. Average Responses to Shocks to Net TaxesUnited States ...............................................4711. Average Cumulative Government Spending and Tax Multipliers ....................................4812 . Baseline Scenario (Adverse Growth Shock, No Fiscal Consolidation) ............................4913. Debt-to-GDP Ratio Following a Smooth Fiscal Consolidation (Scenario I) .....................5014. Debt-to-GDP Ratio Following a Frontloaded Fiscal Consolidation (Scenario II) ............51 15. Debt-to-GDP RatioDifference Between Frontloaded and Smooth Fiscal .....................5216. Debt-To-Gdp RatioDifference Between Frontloaded (Spending-Cuts-Only Based) ....5217. Debt-to-GDP RatioDifference Between Frontloaded and Smooth Fiscal .....................5318 . Debt-to-Gdp RatioDifference Between Frontloaded and Smooth Fiscal .....................53

7/28/2019 Wp 12190

5/61

4

I. INTRODUCTION

With the U.S., Europes and Japans public debt at historic levels, concerns are rising

over the growth impact of needed fiscal adjustment. The severe recession, the significant

capital interventions in the financial markets and the fiscal stimulus measures have pushed up

the joint public debt to levels not seen since the end of World War II. Reducing the publicdebt ratio to more comfortable levels would require a large and sustained adjustment that

could most likely weaken aggregate demand. In the United States and Japan, the

conventional monetary policy support to fiscal consolidation is limited by the fact that

interest rates are near their effective floor, while everywhere population aging and currently

low trend growth provide little room to absorb falling demand. As most of the advanced

countries are called to adjust, external demand will provide very little support to U.S.,

Europes and Japans growth prospects.

On the other hand, consolidation cannot be postponed indefinitely. In the United

States, the debt ceiling deal reached just before August 2, 2011 followed a painful

negotiation and was quickly followed by a downgrade of the U.S. credit rating by a major

rating agency. And although, unlike in some countries in Europe, the United States is

currently not facing market pressures to signal immediate consolidation plans, these could

mount eventually, now that the 12-member super committee tasked with finding $1.2 trillion

in deficit-reduction measures failed in delivering a viable solution by late Fall 2011, and the

federal government is again expected to near its new borrowing limit in early 2013. In

Europe, especially in countries where sovereign debts have increased sharply due to bank

bailouts, a crisis of confidence has emerged with the widening of bond yield spreads and risk

insurance on credit default swaps between these countries and other euro zone members,

most importantly Germany. While the sovereign debt increases have been most pronounced

in only a few euro zone countries (notably in Greece, Ireland and Portugal, and more recently

Spain and Italy), they have become a perceived problem for the area as a whole because of

the potentially severe contagion effects.

Thus, if consolidations are delayed there is a real risk of debt downgrades or defaults.

But frontloading consolidation risks bringing recoveries to a halt, hindering the same fiscal

adjustment or making it too costly in terms of jobs and output. One immediate question then

is: what is the pace of fiscal consolidation in the United States, the euro area and Japan that

would achieve maximum adjustment given low growth, while preserving the recovery?

To answer this question it is necessary to estimate fiscal multipliers for various stages

of the business cycle. It is also important to account for the fact that a too-abrupt

consolidation that begins while the economy is expanding can push the economy into a

recession. In other words, the success of a fiscal consolidation (in absolute and GDP terms)

ultimately depends on how the consolidation affects the economic cycle given a set of

7/28/2019 Wp 12190

6/61

5

specific starting conditions. For this we need an empirical methodology that makes the stages

of the business cycle endogenous to the computation of the fiscal multipliers; and we need a

tool to quantify the likelihood that a fiscal consolidation (of a given size) occurring in a

certain stage of the cycle drives the economy into another stage.

Unfortunately, most of the literature of fiscal multipliers is based on linear structuralvector auto regressions (SVARs) or linearized dynamic stochastic general equilibrium

(DSGE) models which by construction rule out state-dependent multipliers. Auerbach and

Gorodnichenko (2012a, AG hereafter), developed a methodology that allows multipliers to

vary between expansion and recession. Building on their framework, we want to study how

fiscal shocks can shift the economy from one regime (expansion, say) to another (recession).

Different from AG and other related papers, we define the regimes in terms of the sign of real

GDP growth and not as output gaps (or as deviations from the trend) as, we believe, the

growth rate better captures the state of the economy. A country, for instance, could be in

better shape when growing out of a negative output gap rather than when contracting after a

period of economic boom. Moreover, the use of the GDP growth sign makes endogenizing

the regimes much easier. Furthermore, we want also to take explicitly into account the role of

the monetary policy stance, which has been indicated as an important additional determinant

of how output responds to fiscal adjustments.

To achieve these objectives, we use a similar methodology to that of AG which yields

results that are richer along three important dimensions:

First, we estimate fiscal multipliers conditional on monetary policy by expanding theregime-dependent VAR with a real short-term interest rate. Traditional VARs used to

estimate fiscal multipliers and rooted in the Blanchard-Perotti (2002, BP

hereafter) framework contain only three endogenous variables, namely real GDP,

public expenditure and tax revenues. However, fiscal multipliers can vary depending

on the stance of monetary policy and, actually, subsequent analyses in the BP

tradition do include variables capturing monetary policy. A recent study by the IMF

(2010) that examined the effects on U.S. output of changes in monetary policy and

tax rates in the United States, for example, shows that reductions in interest rates

usually support output during episodes of fiscal consolidation. Central banks offset

some of the contractionary pressures by cutting policy interest rates, and longer-term

rates also typically decline, cushioning the impact on consumption and investment.And if interest rates are near their effective floor, the effects of fiscal consolidation

are more costly in terms of lost output. Theoretical work by Hall (2009), Woodford

(2011) and Christiano et al. (2011) and empirical work by Ahrend et al. (2006) and

Canova and Pappa (2011) also indicate that the monetary policy stance is an

7/28/2019 Wp 12190

7/61

7/28/2019 Wp 12190

8/61

7

burdens, all which differ markedly in our sample group of countries. We plan to expand the

estimation to other countries in future research as data becomes available.

The multipliers that we find are broadly in line with Auerbach and Gorodnichenko

(2012a, 2012b) prior empirical estimates, and reaffirm their finding that fiscal multipliers

tend to be considerably larger in recessions than in expansions. Importantly, recent researchby Caldara and Kamps (2012) shows that the identification scheme we employ to disentangle

nondiscretionary policy responses to output from discretionary policy choices, based on the

classical approach of Blanchard and Perotti (2002), imply fiscal output elasticities consistent

with plausible auxiliary estimates. Alongside, our estimates of the government spending

multiplier in recessions and expansions are largely consistent with the theoretical arguments

in both (old) Keynesian and (new) modern business cycle models.

The main findings from the analysis can be summarized as follows:

1. Fiscal expenditure multipliers are significantly larger in downturns than in upturns;5

2. While it is plausible to conjecture that confidence effects have been at play in our

sample of consolidations, during downturns they do not seem to have ever been strong

enough to make the consolidations expansionary at least in the short run;

3. Expenditure multipliers (where expenditure is defined as public consumption and

investment only) are significantly larger than tax multipliers (where tax is defined as tax

minus transfers) in downturns;

4. Monetary policy does not seem to have a strong cushioning effect on economic

activity against fiscal withdrawals implemented during downturnspossibly reflecting the

fact that during the actual downturns experienced by the countries of our sample, over our

estimation horizon, interest rates may not have been cut sufficiently (or cut sufficiently fast)

to counteract the drop in output that accompanied the episodes of fiscal consolidation.6 The

weak cushioning effect of monetary policy may also be due to the fact that some of the

downturns in our sample might actually be induced by the monetary authorities in an effort to

lower inflation;

5 When referring to upturns and downturns, we refer to the sign of the growth rate of real GDP. In our analysis.

we abstract from the sign and size of the output gap.

6 Zero bounds on nominal rates may not fully explain this result in that when at the zero bound, central banks in

Europe, Japan and the United States have generally made use of quantitative and qualitative easing measures

which have the potential effect of lowering the real interest rate at various maturities.

7/28/2019 Wp 12190

9/61

8

5. The probability that a fiscal consolidation initiated in a downturn deepens or extends

the downturn is almost twice as large as the probability that a consolidation started in an

upturn triggers a downturn;

6. Strong (defined as 2 standard deviation fiscal shocks) consolidations are 20 percent

more likely to trigger or extend downturns than mild (defined as 1 standard deviation fiscalshocks) consolidations. In other words, the same fiscal adjustment is less recessionary if

made via an extended adjustment as opposed to a more abrupt one;

7. The exact size of the 1-year cumulative fiscal multiplier is country-, time-, and

circumstance-specific, with ranges in our sample countries (in downturns) between 1.6 and

2.6 for expenditure shocks, and 0.16 and 0.35 for tax shocks.

8. The peak effect on output of fiscal consolidations is within the first year from the

shock.

9. Frontloaded consolidations tend to be more contractionary and, hence, delay the

reduction in the debt-to-GDP ratio relative to smoother consolidations.

The key policy implications from the analysis are thus:

Implementing fiscal consolidations during periods of positive output growth reducessignificantly the impact on output;

If consolidations need to be implemented during downturns (for instance, to regainmarket confidence), they should prioritize increases in net taxes (defined as taxes

minus transfers);

If consolidations have to occur during downturns and prioritize cuts in publicconsumption and investment, they should be smooth and gradual and be accompanied

by increases in net taxes. When the fiscal adjustment relies more heavily on net tax

increases than on expenditure cuts, and is gradual, the debt-to-GDP ratio falls by

more, while affecting output less adversely;

Monetary policy should be used more proactively to mitigate the output costs ofconsolidations;

Measures that improve the credibility and durability of consolidations may boostpositive confidence effects, alleviating the cost of future consolidations. In theEuropean context, for example, a strong commitment to a responsible implementation

of the European Unions Fiscal Compact may be warranted in this regard;

Last but not least, more empirical research is needed to understand the size andregime-dependency of multipliers of subcomponents of expenditure, since it is likely

that some expenditure cuts are more output-costly than others (our estimates suggest

7/28/2019 Wp 12190

10/61

9

that, historically, the kind of expenditure cuts used were costly in terms of foregone

output);7 moreover, our analysis abstracts from the role of debt levels, which have

been shown to significantly reduce growth, especially when they exceeds a specific

threshold, usually estimated around 90 percent of GDP (see Kumar and Woo, 2011,

and Baum, Checherita and Rother, 2012). Extending the research in this direction

could give us further information on the appropriate policy mix in highly-indebted

countries and, indirectly, in situations where fiscal consolidation is implemented in

conditions of near-default.

The paper is organized as follows. Section II provides a brief overview of the

literature on fiscal multipliers. Section III describes the methodology used. Section IV

presents the results and contrasts them with findings in the literature, focusing in particular

on earlier IMF results on multipliers. Section V uses estimated multipliers and probabilities

to compute the likely evolution of the debt-to-GDP ratios in the sample countries under

planned fiscal adjustments. Concluding remarks and policy implications follow. A

description of the data employed in the analysis is appended to the paper.

II. LITERATURE REVIEW

Fiscal multipliers have long been the object of theoretical and empirical

investigations. Our analysis relates directly to the empirical literature on the effects of fiscal

shocks (and thus can be used to check the predictions of models like DSGE ones) and

indirectly to at least another two avenues of research on multipliers, namely the literature on

the effects of fiscal policy in a DSGE context, and the analysis of fiscal policy at special

times. Each of these literatures is extensive, and is only briefly summarized below.

A. Empirical Literature on The Effects of Fiscal Shocks

The early empirical literature on fiscal multiplier dates back to the late 1990s.8 At

least qualitatively, these studies generally reach the same conclusions and support the

neoclassical business cycle view of the impact of fiscal policy: in response to a discretionary

positive government spending shock, output increases, consumption and wages decline, non-

residential investment rises, while residential investment falls.

7

By the same reasoning one cannot exclude that cuts to some other subcomponents of public expenditure areactually expansionary, in which case one could infer from our analyisis that these subcomponents were neverthe subject of fiscal consolidation, or else, they were part of consolidations dominated by cuts to subcomponentsof expenditure that are costly to cut in terms of output.

8 Among others, this includes seminal works by Ramey and Shapiro (1998) and Edelberg et al. (1999) whichemploy a narrative ore dummy-variable approach to identify discretionary fiscal policy shocks in a univariatecontext or in vector autoregressions (VAR).

7/28/2019 Wp 12190

11/61

10

However, more recent empirical studies, starting from Fats and Mihov (2001) and

Blanchard and Perotti (2002), adopt structural VARs (SVAR) for the purpose of

identification, and obtain results more in line with Keynesian claims on the impact of fiscal

policy.9 Under the BP approach government spending shocks affect output, consumption and

the real wage positively, while they tend to have a negative effect on investment.10 In

addition, Perotti (2007) and Monacelli and Perotti (2008), propose a variant to the narrative

approach that yields similar results to the SVAR literature. In a cross-country context, work

by Ilzetzki, Mendoza and Vegh (2010) find that multipliers tend to be larger in industrial than

in developing countries; the multiplier for countries with a flexible exchange regime or high-

debt countries is close to zero; the multiplier in open economies is typically smaller than in

closed economies; and finally, the multiplier on government investment is similar to the

multiplier for government consumption. Krichner, Cimadomo, Hauptmeier (2010) (in line

with Perotti, 2004) find that the impact of government spending on output has been

increasing until the 1980s and falling thereafter, showing that the impact significantly

depends on the GDP ratio of household credit and the composition of spending.

In general, quantitative estimates of the multiplier vary widely depending on the

assumptions and techniques used.11 These include: (i) the sample used in estimation; (ii) the

estimation technique; (iii) whether the measuring accounts of automatic stabilizers or not;

(iv) whether the economy is going through a particular phase of the business cycle times

(expansion or recession, high unemployment); or (v) whether spending (or spending

withdrawal) is anticipated or not.

The main results can be summarized as follows:

Low estimates of the output multiplier (below one) can be found in Barro and Redlick

(2011), and IMF (2010), among others. Ramey (2009), and Hall (2009), among others,

9 Fragetta and Melina (2011) support the assumptions underlying the Blanchard-Perotti approach by using a

graphical modeling approach. Ramey (2009) criticizes the Blanchard-Perotti approach on the grounds that it

fails to take into account anticipation effects of fiscal policy. Mertens and Ravn (2011)on one hand show that

anticipation effects may invalidate SVAR estimates of impulse responses; on the other hand they fail to find

evidence that anticipation effects overturn the existing findings from the fiscal SVAR literature.However, as

also Auerbach and Gorodnichenko (2010) argue, the narrative approach imposes a constraint on its own, i.e.

that the effects of only a very specific class of shocks can be evaluated, such as military spending build-ups and

tax changes unrelated to short-term considerations.

10 Pappa (2009), for example, finds that the real wage increases following an increase in government spending.

11 Spilimbergo et al. (2009) and Hebous (2009) provide excellent summaries of findings on fiscal multipliers

reviewing literature up to 2009.

7/28/2019 Wp 12190

12/61

11

estimate multipliers around one; while Blanchard and Perotti (2002), Monacelli et al. (2010),

Blinder and Zandi (2010), Acconcia et al. (2011), Fragetta and Melina (2011), among others,

report values above one. Auerbach and Gorodnichenko (2012a) study asymmetries in the

propagation of fiscal shocks in booms and downturns and report output fiscal multipliers of

up to 2.5 during recessions. Auerbach and Gorodnichenko (2012b) estimate spending

multipliers for a lage number of OECD countries, and systematically relying on real-time

forecast data to purge policy innovations of their predictable component; their results confirm

those of their previous paper. IMF (2012) also employ a regime-switching VAR to control

for differences in the impacts of fiscal shocks during periods of positive and negative output

gaps and find multipliers of up to 1.3 during downturns. Mountford and Uhlig (2009) find

large multipliers (up to 5) for deficit-financed tax-cuts using Uhlig (2005)'s sign restrictions

method to identify a government revenue shock as well as a government spending shock,

when explicitly allowing for the possibility of announcement effects, i.e., that a current fiscal

policy shock changes fiscal policy variables in the future, but not at present. Corsetti, Meier

and Mueller (2010) analyze the determinants of the fiscal multiplier confirming AGs andour finding that the multiplier greatly depends on the economic environment at the time of

the fiscal shock.

Finally, a few studies have focused systematically on the outcomes of fiscal

retrenchments. Most old and new studies conclude that consolidations are not expansionary

in the short run. For example, a recent IMF study (IMF, 2010) focusing on the history of

fiscal retrenchment in 15 advanced economies over the past 30 years, finds that on average a

cut of 1 percent of GDP typically reduces GDP by percent within two years (i.e. a

spending multiplier of ). Results in Coenen et al (2008), and Forni at al.(2010), also lead to

the conclusion that debt consolidations typically have a contractionary effect on output in theshort run, (Afonso, 2010, however, shows results that provide support to the opposite view).

However, starting with Giavazzi and Pagano (1990) and Bertola and Drazen (1991), a

smaller stream of studies focusing on a handful of country cases finds that consolidations can

be expansionary: if consolidations are interpreted as a signal that the share of government

spending in GDP is being permanently reduced, the private sector may revise upwards its

estimate of its permanent income, raising current and planned consumption. Recent work

reappraising the impact of confidence effects on the sign and size of fiscal multipliers

supports the view that fiscal contractions can be expansionary in the presence of strong

expectational effects. Inspired by work from Corsetti et al. (2009, 2010), for example,Cimadomo, Hauptmeier and Sola (2012) show, estimating a structural (non-cycle-dependent)

VAR on U.S. data and the Ramey (2011)s time series of military build-ups to measure

exogenous spending shocks, that expectations about the future fiscal policy stance play a key

7/28/2019 Wp 12190

13/61

12

role in shaping the macroeconomic responses to fiscal shocks. As a consequence, a fiscal

contraction may turn out to be expansionary if the expectation channel becomes sufficiently

strong.12

Why are results on multipliers so dispersed in the literature? Caldara and Kamps

(2008) demonstrate that, once differences in specification of the reduced-form VAR modelare accounted for, all identification approaches used in the literature yield qualitatively and

quantitatively very similar results as regards the effects of government spending shocks.

Caldara and Kamps (2012) extend their earlier analysis to showusing U.S. data-- that

differences in priors on elasticities implicit in alternative identification schemes translate into

large differences in fiscal multipliers. Some of the identification schemes appear very

dogmatic, selecting a single value of the relevant output elasticity. Others appear quite loose,

imposing almost no restriction on the relevant elasticity. They then survey the existing

literature on automatic stabilizers to derive distributions on elasticities that encompass the

existing empirical evidence and estimate fiscal multipliers based on these prior distributions.They conclude that there is no evidence to support the view that tax policy provides a larger

stimulus to output than spending policyfor plausible priors on output elasticities of fiscal

variables, tax multipliers are smaller than spending multipliers.13

B. The Effects of Fiscal Policy: What do Models Say?

In the traditional Keynesian model, fiscal shocks were considered the main tool to

stabilize output. For the shock to have a real impact it had to be deficit financed, otherwise it

did not have an impact on the income of economic agents. The main mechanism is that of the

fiscal multiplier, according to which the increase in income that follows the fiscal shocks ismuch higher than the size of the initial shock because part of it (depending on the marginal

propensity to consume or invest) is used to further increase aggregate demand through

consumption or investment. Spending shocks were thought to have a bigger impact because

the multiplier operated after a one-to-one increase in aggregate demand. Tax shocks had a

12 See also Perotti (2011).

13 More specifically, Caldara and Kamps (2012) demonstrate that empirical estimates of fiscal multipliers

obtained via SVARs that identify fiscal shocks using a BP or a BP-like recursive approach (like we do) arelikely to produce larger spending multipliers than tax multipliers. This is because this identification approach

(correctly) implies a smaller (more plausible) elasticity of output to taxes (around 1.5) compared to e.g. the

narrative or the sign-restriction approach (where the implied elasticity is 3 or larger)which in turn affects

univocally the relative size of spending and tax multipliers in this direction.

7/28/2019 Wp 12190

14/61

13

slightly smaller impact because the multiplier operated on an increase in demand that was

proportional to the marginal propensity to consume.

The key tenet of the neoclassical real business cycle doctrine is that deficit-financed

fiscal shocks (e.g. a fiscal expansion) generate a small positive multiplier (i.e. output changes

less than proportionally to the shock measured in real terms). This is because a positive

government spending shock increases the present discounted value of tax payments and this

triggers a negative wealth effect that dampens consumption, fosters labor supply and curbs

real wages (e.g. Baxter and King, 1993). Tax shocks do not have any impact on output

because, keeping spending constant, a change in current taxes will need to be offset by

future, higher taxes, which leaves the present discounted value of taxes unchanged (the so-

called Ricardian equivalence). The introduction of frictions, however, can change the size

and sign of the impact. Linnemann (2006) demonstrates that with a no additively separable

utility function and a small intertemporal consumption elasticity, higher fiscal spending can

raise consumption and lower investment, consistently with the results of SVAR estimationbased on the BP methodology. Callegari (2007) shows that the introducing of borrowing

constraints in an otherwise standard RBC model can generate positive (albeit small)

multipliers to tax shocks and while amplifying the negative wealth effect that follows

positive spending shocks. Leeper et al. (2010) introduces implementation delays in the

analysis of government investments shocks in a RBC model and find thatimplementation

delays can produce small or even negative labor and output responses in the short run.

Based on the same fundamental assumptions of RBC model, new-Keynesian (NK)

models introduced real and nominal rigidities in a dynamic and stochastic general

equilibrium model. In the absence of additional frictions, the predictions of NK and RBCmodels on the impact of fiscal policy shocks were not so dissimilar, because the wealth effect

tended to prevail over the other propagation channels. The general finding is that the

multiplier of a (temporary) increase in government expenditure is much less than one (see for

example, Coenen and Straub, 2008; Cogan et al., 2010; and Cwik and Wieland, 2009). The

introduction of further frictions into the model can amplify the impact of fiscal shocks. The

introduction of myopic, rule-of-thumb consumers increases the multiplier because limits the

share of agents on which the negative wealth effect works. Given the lack of forward-looking

behavior and a Unitarian marginal propensity of consumption, rule-of-thumb consumers

respond quite strongly to changes in taxes, especially if they are lump-sum. Gal, Lopez-

Salido and Valles (2007), however, show that, in order to generate multipliers in line with theSVAR literature, one also needs to introduce labor market frictions, like unionized wage

bargaining. This dampens the drop of real wage that follows the increase in labor supply of

forward-looking agents, which would compress the income of rule-of-thumb consumers.

Ravn, Schmitt-Grohe and Uribe (2007) show that the presence of deep habits can increase

the spending multiplier given the positive response of consumption, and Cantore, Levine,

7/28/2019 Wp 12190

15/61

14

Melina and Yang (2012) manage to match a series of empirical regularities in the SVAR

literature in a DSGE model with unemployment and deep habit formation.

The predictions of the theoretical models changes quite substantially depending on

whether we are in an open- or in a closed-economy set up, and whether the exchange rate is

fixed or not. In an open economy with a fixed exchange rate, the fiscal multiplier is generallybigger, because of the smaller offsetting wealth effect due to the peg of the interest rate to its

world value. With a flexible exchange rate, however, the impact might be smaller (and even

negative for tax cuts) because of the real appreciation that follows increase in aggregate

demand.

The introduction of distortionary taxation generally reduced the impact of spending

shocks (because of the increased distortion triggered by higher future taxes). Concerning tax

cuts, it all depends on the policy mix; a reduction in lump-sum taxes rebalanced by future

higher distortionary taxes, for instance, can generate a negative output response or greatly

reduce the size of the positive multiplier.

Recent work by Hall (2009) and Woodford (2011) tries to establish the conditions

under which the multiplier is large in simple models. Woodford shows that the government

spending multiplier is (i) necessarily below one in a neoclassical Real Business Cycle (RBC)

model and exactly the same both in an RBC with monopolistic competition and in a sticky-

price NK model with strict inflation targeting; (ii) exactly one in an NK model with fixed real

interest rate; (iii) somewhere between the two values in a model featuring a Taylor rule. In

general, the more accommodative the monetary policy, the higher the fiscal multiplier.

Moreover, substantially larger-than-one multipliers can be obtained in standard NK models if

the ZLB binds (Woodford, 2011).

The way the labor market is modeled plays a crucial role in the propagation

mechanism of fiscal shocks. Monacelli et al. (2011) simulate the effects of fiscal policy on

the labor market using a standard RBC model with search-match frictions and show that the

model largely fails in reproducing the size of the output multiplier whereas it can produce a

realistic unemployment multiplier but only under a special parameterization. Extending the

model to embed New Keynesian features (e.g. unemployment benefits, real wage rigidity

and/or debt financing with distortionary taxation) only worsens the picture. When

complementarity is coupled with price stickiness, however, the magnification effect can be

large. Likewise, Cantore et al. (2011) manage to match a series of empirical regularities in

the SVAR literature in a DSGE model with unemployment and deep habit formation.

7/28/2019 Wp 12190

16/61

15

C. Fiscal Policy Under Special Conditions

Most papers that focus on normal times find multipliers to be close or below one.

Hall (2009) estimates a spending multiplier of around one for the United States; while Barro

and Redlick (2011) estimates of the spending multiplier are in the 0.60.7 range. Ramey

(2011) finds a multiplier of military spending 0.61.1 (controlling for potential anticipationeffects). Not controlling for anticipation effects, the multiplier drops to 0.5. Zandi and

Blinder (2010) find a general multiplier of 1.5 (ranging from 1.13 for spending on energy

assistance to 1.7 for spending on food stamps). While Blanchard and Perotti (2002), Perotti

(2007), Canova and Pappa (2007), Montford and Uhlig (2009) (VARs) estimates of the

impact multiplier all tend to fall between 0.4 and 1 (although occasionally a larger multiplier

is obtained). Corsetti et al (2011) also find estimate point estimates of the multiplier in the

range 1.21.4 on impact with confidence intervals however including 1 using data on

spending in infrastructure at provincial level in Italy, and an instrument identifying changes

that are large and exogenous to local cyclical conditions.

By contrast, most papers that focus on special times find multipliers to be at or

above 1, especially when the focus is on periods similar to the current one. Barro and Redlick

(2011), for instance, show that the fiscal multiplier of government spending onto output

when unemployment rate is over 12 percent is about 1. Fishback and Kachanovskaya (2010),

focusing on the period of the New Deal when the unemployment rate was between 9 and

25 percent, find that the personal income multiplier ranges from 0.91 to 1.67 depending on

which part of government spending is increased. They also find that multipliers for a more

production-based measure of state income-per-capita is about 10 to 15 percent smaller; and

that the impact of the federal spending on employment is negligible (or even negative).

Romer and Burstein (2009) estimate a multiplier during the global financial crisis at

about 3 or even largera finding that validates model results in Corsetti, Meier and Muller

(2010) who show that multipliers tend to be larger during financial crises. Auerbach and

Gorodnichenko (2012b) estimate government purchase multipliers for a large number of

OECD countries, allowing these multipliers to vary smoothly according to the state of the

economy and using real-time forecast data to purge policy innovations of their predictable

components. They adapt their previous methodology (Auerbach and Gorodnichenko, 2010)

to use direct projections rather than the SVAR approach to estimate multipliers, to economize

on degrees of freedom and to relax the assumptions on impulse response functions imposedby the SVAR method. Their findings confirm those of their earlier paper. In particular, GDP

multipliers of government purchases are larger in recession, and controlling for real-time

predictions of government purchases tends to increase the estimated multipliers of

7/28/2019 Wp 12190

17/61

16

government purchases in recession. Multipliers are found to vary also when other economic

conditions vary beyond the cycle or the state of the financial sector.14

Another debate surrounds whether the spending multiplier is larger when monetary

policy is highly accommodative, like now that the monetary policy rate of many central

banks is at its effective lower bound. Both Christiano et al. (2011) and Woodford (2011)demonstratein a theoretical contextthat when interest rates are at their effective low, like

during the recent global financial crisis, fiscal shocks tend to have magnified effects (with

multipliers as large as 10) because governments spending cannot crowd out private spending,

notably business capital expenditure.

III. ECONOMETRIC METHODOLOGY

The methodology that we use is an adaptation of the approach proposed by Balke

(2000). Similar approaches have recently been adopted also by Calza and Sousa (2006),

Afonso et al. (2011) and Baum and Koester (2011). We follow four steps. First, we test for

and estimate a threshold vector-autoregressive model that changes structure if growth

crosses a critical threshold. Here, growth changes are endogenous in the sense that shocks

such as government expenditure or tax revenue can result in a switch between different

regimes of the business cycle. Second, using nonlinear impulse-response analysis, we isolate

the relative effects of shocks and the nonlinear structure on the time-series behavior of

output. Third, we compute regime-dependent multipliers. Finally, we run stochastic

simulations to compute the probabilities of regime switching. These steps are described in

more detail below.

A. The Model

We specify a threshold vector-autoregressive (TVAR) model with endogenous

regimes. The TVAR model is desirable on a number of grounds. First, it allows us to capture

possible nonlinearities such as asymmetric reactions to shocks in a simple way. Because the

effects of the shocks are allowed to depend on the size and the sign of the shock, and also on

the initial conditions, the impulse response functions are no longer linear, and it is possible to

distinguish, for instance, between the effects of fiscal developments during an expansionary

14 Gal et al. (2007) quantify the multiplier at 1.9 if share of rule-of-thumb consumers is 0.75 (such high

proportion of rule-of-thumb consumers could be interpreted as a consequence of the fact that a large share of

households have no access to financial markets or decide not to participate in them). Uhlig (2010) shows that

the way in which government expenditure is financed matters for the size of multipliers. If debt is contained

increases in distorting taxes could lead to a negative multiplier. Canova and Pappa (2006) show that negative

multipliers emerge in U.S. states with strict balance budget requirements.

7/28/2019 Wp 12190

18/61

17

phase and those during a recessionary phase. Second, in a TVAR shocks can trigger switches

between one regime (e.g. expansion) and another (e.g. recession) since the threshold variable

is in turn a function of one of the endogenous variables.

Our structural TVAR can be expressed as follows:

(1)

Where is a vector containing total government expenditures, realoutput, net taxes (all in logreal per capita terms) and the short-term real interest rate,respectively. So our model includes the classical Blanchard-Perotti variables but also a

synthetic measure of the ex-postmonetary stance to condition on the non-fiscal component of

the policy mix. Using taxes net of transfers to compute tax multipliers not only keeps with

the bulk of the literature, but it also makes economic sense in that tax and transfers should

have similar multiplicative effects on activity, as they affect disposable income in a similar

waywhile they plausibly have a rather different impact on activity than public consumption

and investment.15 and are lag polynomial matrices, while are structuraldisturbances. Variable is the threshold that determines which regime the system is in,and is an indicator function that equals 1 when and 0 otherwise. Thethreshold variable, , is real output growthdefined as log-difference, 100 i.e. a function of real output itself (which, in turn, is an element of). At least one lag of thethreshold variable is needed in order it to recursively feed into the VAR dynamics. We thus

set 1.16 In addition, we set the critical value for the threshold, , equal to zero in order todistinguish clearly between periods of positive growth and periods of negative growth.

However, (i) we test for this restriction and (ii) if we estimate the threshold itself with somevariants of a Wald test in a grid-search fashion, as proposed by Balke (2000), we obtain a

15 In principle social security contributions could have a different GDP impact than personal income taxes ortransfers (because households perceive a link between payroll and pension benefits) but this effect is likely to besmall. In addition, inasmuch transfers are not well or perfectly targeted, their effect on GDP may be smallerthan the effect of taxes, which would diminish the augmented effect that transfers may have over an abovetaxes, considering perceived links between payroll and pension benefits.

16

Generally dis an unknown parameter, and Hansen (1997) proposes a single-equation estimation procedurethat allows to estimate it besides the other unknown parameters of a TVAR model. Extending this procedure tothe multivariate case may in principle lead to a longer delay with which regime-switches affect VAR dynamics.However, we opt for following related studies (see e.g. Balke, 2000; Calza and Sousa, 2006; and Afonso et al.,2011) and setting the delay to one quarter as here we are not interested in the response to fiscal shocks when aregime switch occurred long ago, but rather when it has just occurred.

7/28/2019 Wp 12190

19/61

18

threshold very close to zero.17 As a result, two regimes govern the dynamics of the TVAR:

output expansions and output contractions. The TVAR describes both the evolution of andthe output growth regimes.

This implies that shocks to total government expenditures, real output, net taxes and

the short-term real interest rate can determine whether the economy is in a positive ornegative output growth regime. In addition to the lag polynomials changing across regimes,

contemporaneous relationships between variables may change as well. In other words and, which reflect the structural contemporaneous relationships in the two regimes, maydiffer. We assume that andhave a recursive structure in a standard Choleski fashion.We use the assumption proposed by Blanchard and Perotti (2002) that government spending

is unable to react to output and other shocks within a quarter due to implementation and

decision lags typical of the budgeting process. If identification is achieved via a Choleski

decomposition, this assumption translates into ordering government spending first. The same

approach to identification has been employed by Monacelli et al. (2010). The variable

ordering is thus: total government expenditures, real output, net taxes and the short-term realinterest rate. While this recursive structure is not without controversy, much of the recent

fiscal VAR literature uses a similar recursive ordering (see Caldara and Kamps, 2006; 2012

for a discussion). In particular, the Choleski approach implies a possibly restrictive zero

impact response of output to net taxes. However, Caldara and Kamps (2012) show that the

output elasticity of net taxes implied by the pure BP approach is only slightly larger than the

elasticity implied by the recursive approach, this entailing that the impact BP tax multiplier is

only slightly larger than zero. Lastly, as we use quarterly data, we impose a lag structure of

4.18

B. Testing for the TVAR Structure

We test for the TVAR structure by imposing the null hypothesis of a linear VAR

against a threshold alternative with 0. The test is conducted by constructing a likelihood-

17 In all cases considered below, this choice allows us to isolate regimes with negative real output growthcontaining at least 20 percent of observations. This satisfies the recommendation made by Hansen (1999) thateach regime should contain at least 10 percent of observations.

18 The selection of the number of lags in this literature is necessarily arbitrary because typically, lag-selection

critiera tend to suggest different lag truncations. In a non-linear context, testing for the optimal lag length is

even more cumbersome. We then chose a specification with 4 lags on the following ground: (i) standard lag-

selection criteria on our data using linear VARs indicate an optimal lag length of 3 to 4 lags; (ii) 4-lag VARs area common choice in the empirical fiscal literature (see, for instance, the many contributions by Perotti) ; (iii) the

consequences of overestimation of the order are less serious than underestimation (Kilian, 2001); (iv) since the

estimation of the VAR coefficients in OLS is done equation by equation even a 4-lag specification does not lead

to severe overparametrization in our case (the parameters to be estimated are 17 in our case, per each country

data set).

7/28/2019 Wp 12190

20/61

19

ratio test statistic for the multivariate case as in Doornik and Hendry (1997). Asmentioned, imposing a zero threshold does not turn out to be a strong assumption as variants

of the Wald test, such as the sup-Wald with methods of inference proposed by Hansen

(1996), select a threshold in the vicinity of zero in all cases considered.

C. Non-Linear Impulse Response Functions

To evaluate the effects of fiscal consolidations and then compute regime-dependent

fiscal multipliers, we conduct impulse-response analysis. The nonlinear structure of the

model makes this task more complex than in the linear case as in TVARs the moving-

average representation is not linear in the disturbances (as some shocks may lead toregime switches), hence the necessity to resort to numerical methods.

The impulse response function at horizon k, , can be thought of as the revisionin the conditional expectation of as a result of knowing the value of an exogenous shock

. In other words, , can be expressed as the difference between the expectation of conditional on a particular history and a particular realization of the shock, i.e.|, , and the analogous conditional expectation in which shocks are purelystochastic, i.e. |.

In order to isolate the effect of a single exogenous shock, say to government

expenditures or net taxes, the value of just one element at a time in is set to a specificvalue while the remaining elements are randomly drawn. It follows that (a) unlike in linear

models, the impulse-response function for the nonlinear model is conditional on (i) the past

history of the variables, (ii) the size, and (iii) the direction of the shocks; and (b) calculating a

nonlinear impulse-response function requires specifying the nature of the shock (its size andsign) and an initial condition. Such computations can only be made by resorting to numerical

simulations. These are carried out by randomly drawing vectors of shocks , 1, , ,and then simulating the model conditional on an initial condition and a given realizationof. As the choice of a starting value would be arbitrary, the procedure takes every datapoint in a given regime as the initial condition and runs 500 stochastic simulations for a

15 quarter horizon starting from each of them. Then, the average of these will represent the

conditional expectation.

D. Regime-Dependent Fiscal Multipliers

Based on the nonlinear impulse responses we can compute regime dependent

multipliers of fiscal consolidations, i.e. the quantitative effect that a fiscal consolidation has

on real output, conditional on taking place in a given regime. As we allow for endogenous

regime switches in the nonlinear impulse responses, these are translated also in the size, sign

and dynamic patterns of implied fiscal multipliers.

7/28/2019 Wp 12190

21/61

20

We report average cumulative multipliers based on negative shocks to government

spending and positive shocks to net tax shocks of size one standard deviation. We use the

definition of cumulative fiscal multiplier reported by Spilimbergo et al. (2009). The authors

argue that this is an appropriate measure of the fiscal multiplier. In fact it summarizes the

effects that a fiscal measure has over a certain time horizon. The cumulative expenditure

consolidation multiplier at horizonN, , is defined as

, where

is the change in output with respect to baselinej periods ahead of the fiscal shock, and is the change in government expenditures at the same time horizon. Analogously, the

cumulative net tax multiplier at horizonN, , is defined as

,

where is the change in net taxesj periods ahead of the shock.

E. Conditional Probabilities of a Recession

Based on stochastic simulations, we compute the ex-ante probabilities of being in arecession regime conditional on starting the simulations in a recession regime or in an

expansion regime. We allow for three alternative policy options: (i) no consolidation;

(ii) mild consolidation (contraction of government expenditures or hike in net taxes by one

standard deviation); and (iii) strong consolidation (contraction of government expenditures

or hike in net taxes by two standard deviations).

In particular, we run 500 stochastic simulations and we save the percentage of cases

in which output contracts at given time horizons. In order to condition on a particular regime

we choose, as starting values of the 500 stochastic simulations, all points in the sample (one

at a time) that fall in that particular regime and then we report the average outcome. In orderto condition on the three alternative policy options we: (i) assume that all shocks are purely

random in the case of no consolidation; (ii) condition on a realization of the government

expenditure shock (or the net tax shock) of minus (plus) one standard deviation for the case

of mild consolidation; and (iii) condition on a realization of the government expenditure

shock (or the net tax shock) of minus (plus) two standard deviations for the case of strong

consolidation.

IV. RESULTS

This Section discusses the results obtained by following the procedure outlined in

Section III. Table 1 reports the results of a likelihood ratio test for the nonlinear VAR

structure. In all cases considered, the null hypothesis of a linear VAR is rejected in favor of

the TVAR specification.

7/28/2019 Wp 12190

22/61

21

A. Non-Linear IRFS

Figures 1-10 report the (non-linear) average impulse responses of output, government

expenditure, net taxes and the short term real interest rate to a government spending shock

for the Euro Area19, France, Italy, Japan and the United States.20

To better highlight the potential implications of the nonlinear structure of the model,

we report results for: (i) the two regimes; (ii) negative and positive shocks (as they are not

necessarily symmetric); and for (iii) a one standard deviation shock and a two standard

deviation shock since results are not necessarily proportional when the scale of the initial

shock is augmented. It follows that, in each figure, the left-hand-side column reports the IRFs

corresponding to the expansion regime; the right-hand-side column reports those relative to

the recession regime; and each subplot contains four lines, which correspond to positive and

negative fiscal shocks that can be as big as one or two standard deviations. The main findings

are the following:

Independently of the state of the economic cycle, fiscal consolidations operated viaspending cuts reduce output in the short run. In both regimes, output contracts

(expands) following a negative (positive) spending shock. Thus, if confidence effects

have been at play in our sample over the short term, they have not been strong enough

to undo the contractionary effects of fiscal withdrawals.

Spending cuts initiated during recessions are contractionary over the entire simulationhorizon.

However, in Japan, the United States and to a lesser extent in the Euro Area spendingcuts that begin during expansions are contractionary only in the short run, becoming

expansionary in the long run. One explanation for this is that agents anticipate a fall

in the tax burden (as government spending has contracted and the economy is

expanding and vice versa)a phenomenon dubbed in the literature confidence

effect. Our empirical estimates thus support the expansionary implications of

confidence effects, but only in the long run, and only when consolidations are

initiated during expansionary phases of the economic cycle.

19 Euro Area here refers to a single unit coming from the aggregation of national data into one single entity.

Please refer to the data appendix for details on the data sources.

20 Strictly speaking, the non-linear impulse responses reported in the figures are averages of stochastic

simulations as explained in Section III.C. Given that the nonlinear procedure does not allow computing proper

confidence intervals as in linear VAR models, we prefer not to report or discuss statistical significance.

7/28/2019 Wp 12190

23/61

22

With the exception of France, the Euro Area and Japan (in the recession regime),fiscal consolidations operated via cuts in spending reduce net taxes in the short run.

Both in recession and in expansion, a spending contraction (expansion) makes net

taxes decrease (increase). This reflects the drop (rise) in tax revenues that follows a

drop (rise) in output. Subsequently, net taxes increase (decrease) but for different

reasons: in a recession tax rates increase (decrease) and/or are accompanied by a(n)

reduction (increase) in transfer payments; in an expansion net taxes increase

(decrease) in line with the medium-term expansionary (contractionary) effect of the

government spending cut (rise).

Consolidations operated via cuts in spending during recessions tend to raise the realinterest rate in the Euro Area, Japan and the United States. During a recession, a

government spending cut (rise) puts downward pressure on prices and inflation,

pushing up the real interest rate. This finding contradicts in part findings from

simulated models (e.g. IMF, 2010) embedding calibrated Taylor-type rules where byassumption the nominal rate falls by more than inflation expectations during a

recession, as the central bank tries to alleviate the negative consequences of a fiscal

consolidation. It also suggests that in practice, monetary policy may not have been

proactive enough during consolidations that originate in a recession. This may be

related to longer-than-expected lags in the effect of monetary policy, but also an

overestimation of inflation and/or an underestimation of the recessionary effects of

fiscal consolidation on economic activity. The weak cushioning effect of monetary

policy may also be due to the fact that some of the downturns in our sample might

actually be induced by the monetary authorities in an effort to lower inflation. By

contrast, during an expansion, monetary policy seems more proactive: an increase

in government spending elicits a stronger reaction of the central bank than in a

recession, likely a result of the central bank to limit the inflationary consequences of

the shock.

In the euro area, Italy and the United States, in response to a tax hike, the real interestrate increases in the expansion regime and falls in the recession regime. This gives an

indication that monetary policy has been somewhat more accommodative during

recessions in response to tax based consolidations.

Finally, for Japan, France and the euro area (in the expansion regime), a tax increaseinitially (and for France persistently) raises output marginally, although for Japan and

the Euro Area the long-run cumulative output effect of tax hikes is negative as

expected and as in other countries in our sample. One reason may be that the tax hike

is accompanied by a rise in government expenditure in the short run that propels

output momentarily.

7/28/2019 Wp 12190

24/61

23

B. Fiscal Multipliers

In Table 2 and Figure 11, we report average cumulative multipliers based on

consolidation shocks of size 1 standard deviation. The key results are as follows:

In all countries, a fiscal consolidation is substantially more contractionary if madeduring a recession than during an expansion. Spending multipliers in recessions are

up to 10 times larger than spending multipliers when economies are expanding.

(Auerbach and Gorodnichenko 2012a, 2012b find even larger ratios, while

IMF, 2012, finds slightly smaller ratios than us).

The exact sizes of 1- and 2-year cumulative fiscal multipliers are country-specific, butcumulative multipliers are rather homogenously-sized across countries. In our sample

of countries, 1-year cumulative multipliers of consolidations begun during downturns

range between 1.6 and 2.6 for expenditure shocks, and 0.16 and 0.35 for tax shocks

(France and Japan present the opposite sign).21 2-year cumulative multipliers have

similar sizes to 1-year multipliers, implying that the large part of the impact of fiscal

shocks on output materializes within 4 quarters. The absolute sizes of these

multipliers are thus broadly in line with IMF (2012) and Auerbach and

Gorodnichenko (2012a, 2012b). Since none of these works controls for the role of

monetary policy in attenuating the business cycle effect of spending shocks (which is

less relevant for tax shocks), finding similar-sized multipliers to those found in these

works is an indication that, historically, real interest rates changes have not been

sufficient to undo the adverse impact of fiscal withdrawal on economic activity. One

reason may be that during and beyond the Great Moderation, the Fed, the ECB andthe Bank of Japan have controlled inflation tightly and thus changes in real interest

rates have tended to be limited even during recessionary episodeseven when

nominal rates have been actually slashed to their effective low in support of economic

activity.

The size of the multipliers is also compatible with the response of the economies to

fiscal stimuli in the countries of estimation. Japan is a case in kind: for several years

since the early 1990s, fiscal stimuli have been used in association with monetary

stimuli to revive growth following the collapse of an asset price bubble and a long

period of deflation. During most of these years (19 out of 22 looking at the

21 Italys multipliers tend to be smaller than in other countries in our sample in line with the view that theconsumption smoothing effect is likely to be strong given that private wealth is very large (Italy has one of thehighest household wealth-to-income ratios among advanced economies).

7/28/2019 Wp 12190

25/61

24

1990 2012 sample), Japan was expanding, albeit modestly, meaning that only the

smaller (expansion) multipliers would have been at play apart from during strictly

recessionary year. And even during recession years, when the larger multipliers

would have been at play these only capture the impact on output of discretionary

fiscal policy: but Japans extraordinary fiscal stimuli may have failed to lead to larger

gains in output also if antithetic shocksfor example a trend real appreciation of the

yen, or a trend increase in net saving from continuous household and corporate

deleveraginghave been at work, undoing the beneficial effects of fiscal expansion.

Finally, it is widely accepted that fiscal policy in Japan will have limited impact on

the economy as long as structural impediments on the supply side remain. And results

are not necessarily at odds with the episode of fiscal expansion in the United States

during the global financial crisis. True, U.S. growth did not experience a visibly

immediate recovery in the four quarters following the 2008-09 fiscal expansionas

our estimated impulse responses would suggest, the size of the stimulus itself is

debated (see Taylor, 2011 and CBO, 2012)when measured in terms of changes inthe cyclically-adjusted balance of the general government. More importantly, the

severity of the recession, low global demand and historically-low confidence in the

United States may have acted as headwinds on growth while the fiscal expansion was

at play, undoing or delaying some of the positive effects of the stimulus .

Expenditure multipliers are significantly larger than tax multipliers in downturns (upto 10 times larger) but less so during expansions (up to 6 times larger)afinding

broadly in line with IMF (2012); and Auerbach and Gorodnichenko (2012a, 2012b).

The finding that spending multipliers are largersometimes considerablythan taxmultipliers already exist in the literature employing linear VARs. This difference is

magnified further when regime switches are taken into account as the spending

multiplier becomes bigger in downturns while the tax multiplier remains small.

However, it is quite difficult to reconcile the whole set of results with a single

theoretical approach. A tax multiplier bigger than zero but smaller than the spending

multiplier seems to conform with the prediction of the traditional Keynesian model.

They can also be reconciled with a New-Keynesian model with rule-of-thumb

consumers and distortionary taxation, where the tax cuts are mostly on lump-sum

payments but the resulting increase in debt is repaid in the future by an increase in

distortionary taxation.

The asymmetric response of the multipliers to the cycle is much more difficult to

reconcile with the existing theoretical models. Canzoneri, Collard, Dellas and Diba

(2011) try to rationalize this result by introducing financial frictions based on the

models of Woodford and Curdia (2009, 2010). They show that allowing for

countercyclical financial frictions (in this case, costly financial intermediation) can

7/28/2019 Wp 12190

26/61

25

generate bigger multipliers during recessions. According to their model, in time of

negative output gaps financial intermediation is less efficient and more costly;

positive spending, then, generates a more-than-proportional impact on GDP thanks to

the reduction in the inefficiencies linked to the intermediation costs.

The asymmetric response to the different stages of the cycle can also be due to an

increase in the share of agents with limited asset participation. Bilbiie, Meier and

Mueller (2008), for example, show that an increase in the share of these agents can

change the propagation mechanism of fiscal shocks increasing the size of the

multiplier.

In line with the existing literature that does not control for the state of the business

cycle when estimating the response of output to fiscal shocks, multipliers estimated on the

same sample of data using linear VARs are generally smaller than multipliers estimated

specifically over downturns (see Table 2), apart from those estimated on Japanese andFrench data which remain very sizeable. However, even on linear estimates, expenditure

multipliers remain in the majority of cases much larger than tax multipliers.

C. Conditional Probabilities of a Recession Following a Fiscal Shock

Based on the stochastic simulations, in Tables 3 and 4, we report the probability of a

recession regime conditional on starting the simulations in a recession regime or an

expansion regime in the case (i) of no consolidation; (ii) of a mild consolidation (1 SD) and

(iii) of a strong consolidation (2 SD).

The results are as follows:

A spending-based consolidation, on average, delays the transition to an expansionregime as it makes a recession more likely in the next quarter, especially if big. In the

first quarter after the shock, the probability of being in the recession regime is

increased by about 20 percentage points in the United States and by about

15 percentage points in Japan and the euro area, if the shock is big.

But a spending-based consolidation policy may also considerably affect theprobability of going into a recession regime if the consolidation occurs in an

expansion regime, especially with a big shock and in the short run. In the baseline

scenario of no consolidation, the probability of being in a recession, obtained by

simulating the model from a typical starting point drawn from the expansion regime,

converges to around 30 percent. In the quarter following the shock, the probability of

being in a recession increases by up to 20 percent with a big shock. In those countries

where we detect confidence effects in the expansion regime, consolidations may

7/28/2019 Wp 12190

27/61

26

make recessions less likely over longer horizons. It must be said that these

simulations yield average scenarios. We may conjecture that if the shock occurred

at a time of positive, though very low growth, the probability of the fiscal

consolidation sending the economy into a recession would be even higher.

Consistent with the finding of small tax multipliers, tax-based consolidationsgenerally do not greatly affect the probabilities of recessions.

V. WHAT DOES THIS MEAN INPRACTICE?

The results of the TVAR described in the previous sections show that an identical

fiscal shock can have quite different impacts on economic activity depending on whether the

economy is in a period of expansion or recession, depending on the nature of the shock

(expenditure versus taxes) and depending on the intensity of the shock. To give a sense of

how this translates into practical prescriptions for the design of fiscal consolidation plans, in

this section, by using the estimated regime-dependent multipliers on Italy, we derive the

impact of two alternative consolidation plans onto the debt-to-GDP ratio: a smooth

consolidation versus a frontloaded (i.e. more aggressive) consolidation plan.22

The simulations (Figures 1217) show that frontloading fiscal consolidations tends to

have harsher and more protracted adverse effects on output, without accelerating the drop in

the debt-to-GDP ratio relative to a more evenly-distributed consolidation. This can be

problematic in periods of low confidence in the sovereign because failure of fiscal

adjustment to have immediate visible effects on debt may further reduce credibility, pushing

further up risk premia on government paper and ultimately making the consolidation more

costly and less effective. Alongside, the simulations indicate that if a frontloadedconsolidation is based primarily (60 percent and over) on cuts to expenditure, the debt-to-

GDP ratio increases relative to the case in which a more gradual consolidation path is

adopted. Therefore, the compositional design of consolidations (in terms of its taxes versus

spending components) is a key determinant of the ultimate change in the debt-to-GDP ratio

both in absolute terms and in relation to smoother consolidations.

Finally, the ability of curbing the debt-to-GDP ratio depends strongly on the size of

risk premia embedded in the interest cost of servicing the public debt. When risk premia

depend on the level of the stock of debt (for example, assuming that every percentage point

22 In this sense, this exercise can be considered the reverse of what Favero and Giavazzi (2012) do. While they

include the flow government budget constraint into a fiscal VAR, we include the estimated IRFs from the

TVAR described above into the government budget constraint.

7/28/2019 Wp 12190

28/61

27

of additional debt in terms of GDP relative to a level of 60 percent augments the risk

premium by 1 basis point), or even when they are exogenous to the level of debt but are

equal or larger 0.6 percentage points per quarter (around 2 percentage points per year),

frontloaded consolidations deliver even worse results in terms of debt-to-GDP ratios relative

to gradual consolidations (Figures 17 and 18).

Hence, the main implication is that, during recessions, smooth consolidations based

primarily on tax measures may deliver a similar, if not better, debt reduction than more

aggressive consolidations, while affecting output less adversely. This is even more true when

the initial conditions of a consolidation are adverse (the stock of debt is elevated and public

debt yields contain a high and/or rising risk premium component). On the other hand,

aggressive consolidations have the ability to deliver swifter improvements in the primary

cyclically-adjusted balance. The merits of an aggressive consolidation versus a gradual one

thus lie in whether market confidence is faster and more steadily secured by improving

rapidly the structural balance, the debt-to-GDP ratio and/or shielding activity from the most

virulent effects of fiscal adjustment.

A. Model Used for The Debt Simulations

The simulations are based on a stylized growth model and a fiscal bloc. Fiscal shocks

to both spending and taxes are assumed to affect output in line with our estimated TVAR

impulse-response functions (IRFs), which in turn are contingent on the state of the business

cycle. For ease of exposition, the model assumes that the impact of anticipated fiscal shocks

is the same as those of unanticipated shocks. This assumption is not harmless: it tends to

underestimate the announcement effect, reducing the initial impact of shocks that are

distributed over time. (During a recession, however, the impact of this assumption might beless strong, given the more binding borrowing constraint that makes economic agents more

sensible to current economic conditions and thus less forward-looking).

The growth model generates a baseline path for GDP growth that abstracts from the

impact of the fiscal policy shock. In the baseline, output growth depends on the assumed path

for potential GDP, some exogenous demand shock and the size of the output gap. An

endogenous feedback rule from interest rates to output closes the model, thereby

accounting for the role of monetary policy as well.

Based on the baseline scenario, we construct an initial scenario which embeds a

gradual fiscal consolidation that curbs output in line with our estimated IRFs. We then shock

this initial scenario with a more frontloaded fiscal consolidation path, combined with a higher

initial level of debt and different assumption on the way the interest rate is assumed to evolve

over time.

7/28/2019 Wp 12190

29/61

28

The baseline growth model is described by the following equation:

1 1

whereyt is the level of output at times t,yP

is the level of potential GDP,gap is the output

gap, i-iRF

is the spread from the risk-free rate and 1 and 2 are the respective elasticities, bothset equal to -0.1. 23dEXis a variable capturing the role of external demand in propelling or

lowering the level of output. Thus dEXaccounts for exogenous shocks that are not fiscally

related and that can temporarily reduce the level of output. The model also includes an

endogenous response of output whenever the interest rates are above its risk-free level, where

the risk-free level is set equal to the rate of potential growth (which we assume constant in

this simulation).24

On the fiscal sector front, the public debt-to-GDP ratio is assumed to evolve in line

with the standard debt dynamics equation:

1

where i is the interest rate,gthe nominal GDP growth rate, dis the debt-to-GDP ratio andpb

is the primary balance-to-GDP ratio. The first term on the right-hand side describes the

contribution of the interest-growth differential to debt accumulation, also indentified as

snowball effect. The second term describes the contribution of the primary balance. 25 The

primary balance, in turn, is the sum of a structural and cyclical component:

with the cyclical component defined as:

23 When 1 takes a negative value the equation above implies that, in absence of other shocks, the output gap is

closed mechanically as GDP grows above its potential level until the gap approaches zero.

24 This last assumption implies that, in a steady state with a zero primary balance and output growing at itspotential rate, the snowball effect (i.e. the difference between the nominal GDP growth and the debt interestrate multiplied by the previous year debt-to-GDP ratio) is zero and the debt-to-GDP ratio remains constant. Thisassumption is made in order to simplify the analysis, and ensure that each deviations of the debt level from itsstarting value is an effect of the shocks hitting the economy and not a consequence of steady-state dynamics Ofcourse, alongside an exogenous (i.e. constant) we could introduce endogenous risk premia (or vice versa), butthis would simply reinforce our result.

25 In this model, we assume zero debt-deficit adjustments.

7/28/2019 Wp 12190

30/61

29

For the purpose of these simulations, a fiscal shock is defined as a change in the

structural primary balance. It can be realized through a change in taxes or in government

spending. In presence of a fiscal consolidation shock, the baseline output level is reduced

each period by the cumulative impact of the shock.26 Formally, this can be expressed as:

11

where timp andgimp are the total tax and spending multipliers, respectively, (coming from

current and past shocks) applied to output on period t. They are the weighted averages of the

impulse responses of the shocks originated during period of recessions and the impulse

responses of the shocks originated during period of expansion, where the weights are given

by the probabilities of recessions and expansions endogenous to the TVAR.

In turn, the two total output multipliers are defined as follows:

1

1

where is the probability of being in the recession regime. Each element in the right-handside of the two equations above represent the contemporaneous or delayed impact of fiscal

shocks initiated in periods of recession (R, identified with the periods with negative output

growth) or expansion (E). They can be then defined as follows:

,

,

26 In other words, the impact on GDP is measured applying the path of the cumulative multipliers on the additional (1stdifferences) of the consolidation path across periods. Let us assume that the total structuralconsolidation targeted by a country amounts to 3 percentage points of GDP. Let us also assume that theconsolidation is distributed equally across the periods, implying an adjustment of 1 percentage point of GDPevery period. This means that, compared to the pre-consolidation baseline, we have a structural primary balanceimprovement of 1 percentage point of GDP in the first period, of 2 percentage points in the second period, andof 3 percentage points of GDP in the third period. In the third period, the impact on GDP would amount to thefirst quarter cumulative multiplier (generated by the 1 percentage point of GDP shock in the third period) plus

the second quarter cumulative multiplier (generated by the 1 percentage point of GDP shock in thesecond period) plus the third quarter cumulative multiplier (generated by the 1 percentage point of GDP shockin the first period). This way we consider as shocks only the changes in the structural balance (not the level ofthe balance itself). As specified in the paper, this assumes that pre-announced shocks will have the same impactas the announced ones.

7/28/2019 Wp 12190

31/61

30

whereJ=E, R, IRFt is the Tx1 vector of impulse responses of shocks on tax or spending in

the state E or R on output and the are the 1xT vector of shocks occurred in the Tperiods

before period t.

B. Debt simulations

This section presents simulations of debt following consolidation shocks relative to a

model economy.

On the macroeconomic side, we assume that the country has a potential quarter-on-

quarter growth rate of 0.4 per cent, consistent with a annual growth rate of about 1.8 percent.

Starting from this baseline scenario, we then assume that the economy is hit by a negative

shock due to a contraction of the exogenous demand component. The contraction is assumed

to generate an output gap of 3 percentage points of GDP in 4 quarters. This type of shock

could be the result of a drop in the countrys external demand or a financial crisis that

restricts credit availability domestically. In line with the output gap feedback rule, following

the adverse shock the economy is expected to accelerate its rate of growth, and gradually

converge to its long-run potential output level and growth rate.

Figure 12 shows the dynamics of GDP growth and how the debt-to-GDP ratio and its

components would evolve in the baseline scenario described above. The demand shock

reduces growth, which contracts for 3 periods. Thereafter the economy resumes growth at a

rate above its potential in order to gradually close the output gap. The adverse output shock

increases the debt-to-GDP ratio through two different channels: the denominator effect,

which increases the snowball effect, and the increase in the primary deficit, due to a larger

negative cyclical balance component.

On the fiscal policy side, we assume that the country has an initial stock of debt equal

to 120 per cent of GDP and a zero primary deficit. The interest rate is assumed equal to its

risk-free benchmark, which implies a zero snowball effect in the absence of shocks. 27

Throughout we use the fiscal multipliers estimated for Italy in this paper. We then consider

two alternative fiscal consolidation scenarios: a smooth consolidation scenario (Scenario I)

and a front-loaded consolidation scenario (Scenario II). These are described below.

Scenario I. Under this scenario the fiscal authorities implement a consolidation

plan including an percentage point of GDP adjustment every quarter for five year

(20 quarters) delivering a total structural consolidation of about 5 percentage points of GDP(seeFigure 13). Let us also assume that the consolidation is equally divided across tax and

spending measures. Because of the model structure, the shock occurring during periods of

contraction propagates (both in the initial and in the following periods) according to the IRFs

27 Thus, as discussed, in the absence of shocks, the debt-to-GDP ratio would remain constant.

7/28/2019 Wp 12190

32/61

31

estimated during a period of output contraction.28 Since the fiscal shock is not too large in

any given period, its impact on output is contained and the economy remains in recession for

a period equal to the length of the recession in the (no-fiscal-shock) baseline scenario

(3 quarters), although with larger contraction rates.29

Scenario II. Let us now imagine a different type of consolidation strategy (scenarioII), where the same amount of total consolidation of Scenario I is now achieved in a shorter

period of time, say 6 quarters. Figure 14 shows that in this case the impact on output is quite

strong and more persistent, as expected given our estimated impulse responses. The output

gap now opens up to a full 5 percentage points and the recession lasts longer